Protein serine/threonine phosphatase-1 dephosphorylates ... - Nature

Oncogene (2006) 25, 3006–3022

& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Protein serine/threonine phosphatase-1 dephosphorylates p53 at Ser-15 and Ser-37 to modulate its transcriptional and apoptotic activities DW-C Li1,2, J-P Liu1, PC Schmid1, R Schlosser1, H Feng2, W-B Liu2, Q Yan2, L Gong2, S-M Sun2, M Deng2 and Y Liu2 1 The Hormel Institute, University of Minnesota, Austin, MN, USA; and 2College of Life Sciences, Hunan Normal University, Changsha, Hunan, China

We have previously demonstrated that the serine/threonine protein phosphatase-1 (PP-1) plays an important role in promoting cell survival. However, the molecular mechanisms by which PP-1 promotes survival remain largely unknown. In the present study, we provide evidence to show that PP-1 can directly dephosphorylate a master regulator of apoptosis, p53, to negatively modulate its transcriptional and apoptotic activities, and thus to promote cell survival. As a transcriptional factor, the function of p53 can be greatly regulated by phosphorylation and dephosphorylation. While the kinases responsible for phosphorylation of the 17 serine/threonine sites have been identified, the dephosphorylation of these sites remains largely unknown. In the present study, we demonstrate that PP-1 can dephosphorylate p53 at Ser15 and Ser-37 through co-immunoprecipitation, in vitro and in vivo dephosphorylation assays, overexpression and silence of the gene encoding the catalytic subunit for PP1. We further show that mutations mimicking constitutive dephosphorylation or phosphorylation of p53 at these sites attenuate or enhance its transcriptional activity, respectively. As a result of the changed p53 activity, expression of the downstream apoptosis-related genes such as bcl-2 and bax is accordingly altered and the apoptotic events are either largely abrogated or enhanced. Thus, our results demonstrate that PP-1 directly dephosphorylates p53, and dephosphorylation of p53 has as important impact on its functions as phosphorylation does. In addition, our results reveal that one of the molecular mechanisms by which PP1 promotes cell survival is to dephosphorylate p53, and thus negatively regulate p53-dependent death pathway. Oncogene (2006) 25, 3006–3022. doi:10.1038/sj.onc.1209334; published online 27 February 2006 Keywords: p53; PP-1; dephosphorylation; apoptosis; RNAi; phosphorylation and signal transduction

Correspondence: Dr DW-C Li, The Hormel Institute, University of Minnesota, 801 16th Avenue NE, Austin, MN 55912, USA. E-mail: [email protected] Received 23 June 2005; revised 11 November 2005; accepted 14 November 2005; published online 27 February 2006

Introduction The reversible phosphorylation and dephosphorylation at the serine and threonine residues on proteins play important roles in regulating gene expression (Hunter and Karin, 1992), cell cycle progression (Pawson, 1995), and apoptosis (Gjertsen and Doskeland, 1995; Franklin and McCubrey, 2000; Klumpp and Krieglstein, 2002). In eukaryotes, dephosphorylation at the serine/ threonine sites is largely executed by four major protein phosphatases: phosphatase-1 (PP-1), phosphatase-2A (PP-2A), phosphatase-2B (PP-2B) and phosphatase-2C (PP-2C) (Cohen, 1989; Mumby and Walter, 1993), although other protein phosphatases, including phosphatase-4 (PP-4), phosphatase-5 (PP-5), phosphatase-6 (PP-6) and phosphatase-7 (PP-7), also contribute to this process (Brewis et al., 1993; Bastians and Ponstingl, 1994; Chen et al., 1994; Chinkers, 1994; Huang and Honkananen, 1998). The majority of intracellular protein phosphatase activity has been attributed to PP-1 and PP-2A (Cohen, 1989). Both phosphatases are important in promoting survival because inhibition of the PP-1 and PP-2A activities by either okadaic acid or calyculin A leads to apoptosis of many different types of cells (Gjertsen and Doskeland, 1995). However, the molecular mechanisms by which PP-1 and PP-2A promote survival remain largely unknown. Here, we present evidence to show that PP-1 can directly dephosphorylate p53 at two important phosphorylation sites to negatively regulate its transcriptional activity and apoptotic ability, and thus promote cell survival. The tumor suppressor p53 is a master regulator of apoptosis in many types of cells (Shen and White, 2001; Bargonetti and Manfredi, 2002; Vousden and Lu, 2002; Manfredi, 2003; Oren, 2003). Loss of functional p53 is sufficient to inactivate the apoptotic machinery in many types of tumor cells as well as nontumor cells (Shen and White, 2001). As a master regulator of apoptosis in different cells, p53 regulates apoptosis by two mechanisms. First, it acts as a transcriptional factor to regulate expression of many genes involved in apoptosis (Miyashita et al., 1994; Buckbinder et al., 1995; Miyashita and Reed, 1995; Wen-Schaub et al., 1995; Israeli et al., 1997; Polyak et al., 1997; Wu et al., 1997; El-Deiry, 1998). Second, it can activate Bax in mitochondria to antagonize the antiapoptotic ability of

PP-1 dephosphorylates p53 at Ser-15 and Ser-37 DW-C Li et al

3007

Bcl-2 and Bcl-XL (Mihara et al., 2003; Schuler and Green, 2005). The transcriptional activity of p53 can be regulated through several post-translational mechanisms, including phosphorylation/dephosphorylation, acetylation/deacetylation, ubiquitylation, sumoylation and glycosylation (Giaccia and Kastan, 1998; Meek, 1998; Bode and Dong, 2004). Among these posttranslational modifications, phosphorylation and dephosphorylation have an important impact on both stability and function of p53 (Giaccia and Kastan, 1998; Meek, 1998; Bode and Dong, 2004). So far, 17 phosphorylation sites have been mapped on p53 and these sites are phosphorylated by multiple kinases (Giaccia and Kastan, 1998; Meek, 1998; Bode and Dong, 2004). Phosphorylation of p53 at different sites has an important impact on its function. For example, phosphorylation at Ser-15 has been shown to disrupt the interaction of p53 with the Mdm2 protein, leading to stabilization and an increase in protein levels and transcriptional activity (Shieh et al., 1997; Tibbetts et al., 1999). Phosphorylation of p53 at Ser-15 and Ser37 also impairs the ability of MDM2 to inhibit p53dependent transactivation (Shieh et al., 1997). In comparison to our understanding of the kinases phosphorylating p53, much less is known about the corresponding phosphatases that dephosphorylate p53. Since the hypophosphorylated p53 interacts with Mdm2, it is conceivable that dephosphorylation of p53 would facilitate its degradation and attenuate its transcriptional activity. Indeed, a recent study has shown that dephosphorylation of p53 at Ser-37 by protein serine/ threonine phosphatase-2A alters the function of p53 in both fibroblast and breast cancer cells (Dohoney et al., 2004). Here, we present evidence to show that the protein serine/threonine phosphatase-1 can dephosphorylate p53 at Ser-15 and Ser-37 in human lens epithelial cells (HLECs). Dephosphorylation of p53 at these sites by PP-1 changes its transcriptional activity in both lens and non-lens cells. Moreover, the mutant p53 imitating constitutive dephosphorylation or phosphorylation at Ser-15, Ser-37 or both sites attenuates or enhances the transcriptional activity and apoptosis, respectively. Thus, our results demonstrate that dephosphorylation of p53 by PP-1 has a strong impact on its function, and, moreover, part of the molecular mechanisms by which PP-1 promotes survival is to negatively regulate p53 functions.

Results Inhibition of PP-1, but not PP-2A, activity by okadaic acid or calyculin A induces apoptosis of HLECs Our previous studies have shown that inhibition of PP-1, but not PP-2A, by okadaic acid induces apoptosis of both rabbit and rat lens epithelial cells (Li et al., 1998, 2001a). To test whether this is true in HLECs, we have treated these cells with okadaic acid. As shown in Figure 1A(a), 90 nM of PP2A-I1 inhibited about 15% of the total phosphatase activity of PP-1 and PP-2A. Since

PP2A-I1 is a PP-2A-specific inhibitor with IC50 ¼ 30 nM, 90 nM of PP2A-I1 should inhibit about 90% of PP-2A activity. When 20 nM okadaic acid was used in the reaction, it inhibited about 17% of the total phosphatase activity (Figure 1A(a)). A mixture of both 90 nM PP2A-I1 and 20 nM okadaic acid inhibited about 20% of the total phosphatase activity (Figure 1A(c)). These results suggest that 20 nM okadaic acid is able to inhibit PP-2A activity. To further demonstrate that 20 nM okadaic acid is able to inhibit PP-2A, HLECs, after grown to 100% confluence, were treated with 20 nM okadaic acid for 3 h in the 24-well cell culture plate. After treatment, the cells were washed with PBS and then immediately lysed in the plate for measuring the inhibition of PP-2A. As shown in Figure 1B, 20 nM okadaic acid is able to inhibit 16.8% of the total phosphatase activity of PP-1 and PP-2A, a level very similar to that from the in vitro assay. Thus, 20 nM okadaic acid can inhibit PP-2A both in vitro and in vivo. Next we examined the effect of inhibition of PP-2A on apoptosis, as shown in Figure 1E; cell flow cytometry analysis revealed that inhibition of PP-2A with 20 nM okadaic acid has little effect on apoptosis. Therefore, inhibition of PP-2A does not induce apoptosis of the HLECs, which is similar to those found in rabbit and rat lens epithelial cells (Li et al., 1998, 2001a). When the concentration of okadaic acid was increased to 100 nM, however, it blocked about 88% of the total phosphatase activity, which was stronger than the inhibition by 6 nM PP1-I2, specific inhibitor of PP-1, with IC50 ¼ 2 nM (Figure 1A(b)). A mixture of both 100 nM okadaic acid and 6 nM PP1-I2 inhibited about 94% of total phosphatase activity (Figure 1A(c)). Treatment of the culture cells with 100 nM okadaic acid for 3 h also inhibited about 80% of the total phosphatase activity (Figure 1B). These results indicated that 100 nM okadaic acid can block PP-1. To further confirm that 100 nM okadaic acid inhibits PP-1 activity in cultured cells, we examined the phosphorylation status of RB after treatment by 0.01% DMSO, 20 nM okadaic acid for 24 h, or 100 nM okadaic acid for 0–24 h. Previous study has revealed that RB is dephosphorylated by PP-1 (Ludlow et al., 1993). As shown in Figure 1C, treatment of HLECs with 100 nM okadaic acid caused RB hyperphosphorylation within 1.5 h. Thus, 100 nM okadaic acid is able to block PP-1 activity in cultured cells. When PP-1 is inhibited, okadaic acid induced substantial apoptosis (Figure 1F). Thus, as in rabbit and rat lens epithelial cells, inhibition of PP-1, but not PP-2A, induces apoptosis of HLECs. To further confirm that the inhibition of PP-1 leads to apoptosis, we used calyculin A, a more effective inhibitor of PP-1, than okadaic acid (Ishihara et al., 1989). Treatment of HLECs with 1–6 nM calyculin A induced concentrationdependent apoptosis, as shown in Figure 1G. When 6 nM calyculin A was used to treat HLECs for 0–24 h, it was found that all the treated cells underwent apoptosis within 36 h (Figure 1H and data not shown). The apoptotic nature under calyculin A treatment was further verified by DNA fragmentation (Figure 1I). Oncogene


3008

Apparently, inhibition of PP-1 by okadaic acid or calyculin A all leads to apoptosis of HLECs. Inhibition of PP-1 enhances p53 phosphorylation at Ser-15 and Ser-37 and also alters expression of Bcl-2 and Bax Our previous work have demonstrated that when PP-1 was inhibited by okadaic acid, the tumor suppressor, p53 and its downstream proapoptotic regulator, Bax, were upregulated several folds (Li et al., 1998). To determine whether HLECs also displayed such

Oncogene

response, we treated these cells with different concentrations of okadaic acid and calyculin A. Western blot analysis of both control and treated samples revealed that the basic level of p53 expression in HLECs was high and, in response to different concentrations of okadaic acid or calyculin A, the total p53 level was only slightly upregulated (panel 1 of Figure 2a and b). In contrast, the phosphorylation status of p53 at Ser-15 and Ser-37 was dramatically enhanced as the concentrations of okadaic acid and calyculin A were increased to block


3009

PP-1 activity (panels 2 and 3 of Figure 2a and b). It is well established that hyperphosphorylation of p53 at Ser-15 and Ser-37 enhances its transcriptional activity (Shieh et al., 1997; Tibbetts et al., 1999; Bakkenist and Kastan, 2003). To test this possibility, we have analysed the expression levels of two Bcl-2 family members, Bcl-2 and Bax, downstream targets of p53. As shown in panel 4 of Figure 2a and b, Bax was substantially upregulated. A quantitative analysis revealed that Bax was upregulated about three- to five-fold when the concentrations of okadaic acid and calyculin A were increased high enough to block PP-1 (Figure 2e and f). Under the same condition, Bcl-2, an antiapoptotic gene negatively regulated by p53, was found downregulated about two- to four-fold (Figure 2e and f). RT–PCR analysis of the mRNA levels for p53, Bax and Bcl-2 confirmed their expression patterns at protein levels and also the regulation of Bax and Bcl-2 by p53 at the mRNA levels (Figure 2c and d). Thus, inhibition of PP-1 in HLECs enhances phosphorylation of p53 at Ser-15 and Ser-37, which modulates expression of the downstream genes. Silence of p53 expression by RNAi substantially attenuates apoptosis induced by okadaic acid and calyculin A To confirm the role of p53 in mediating apoptosis induced through inhibition of PP-1 by okadaic acid and calyculin A, we have conducted RNAi to silence expression of the endogenous p53. First, a wild type of siRNA oligo and also a corresponding mutant (control) oligo were synthesized (Figure 3a). These oligos have been previously used by other researchers to silence p53 expression in other human cells (Scacheri et al., 2004). When the two oligos were separately introduced into HLECs for 72 h, silence of p53 expression by the wildtype siRNA oligo reached a level of 88% in comparison with the mutated oligo and a similar level when compared with the nontransfected HLECs (Figure 3b).

When expression of p53 was largely silenced, apoptosis induced by okadaic acid (Figure 3c) and calyculin A (Figure 3d) was substantially inhibited. PP-1cas indirectly interacts with p53 in HLECs The result that inhibition of PP-1 activity by okadaic acid or calyculin A induced dramatic hyperphosphorylation of p53 at Ser-15 and Ser-37 suggests that PP-1 may directly dephosphorylate p53. To explore this possibility, we have performed reciprocal immunoprecipitation-linked Western blot analysis with antibodies against p53 and the catalytic subunit of PP-1 (PP-1cas) in the absence or presence of 100 nM okadaic acid, which help to determine whether p53 and PP-1 interaction only occur under stress conditions. As shown in Figure 4a, the immunoprecipitated proteins by anti-p53 antibody contained both PP-1cas and p53. Analysis of the supernatant after immunoprecipitation by p53 revealed that about 80% of total p53 and 38% of total PP-1cas were brought down by anti-p53 antibody (Figure 4a). Similarly, the immunoprecipitated proteins by anti-PP1cas antibody also contained these proteins (Figure 4b). Under this condition, about 40% of total p53 and 78% PP-1cas were brought down by anti-PP-1cas antibody (data not shown). Immunoprecipitation with normal rabbit serum precipitated neither p53 nor PP-1cas (Figure 4c). To further explore the interaction between p53 and PP-1cas, an in vitro binding assay was conducted and the results revealed that PP-1cas does not seem to bind to p53 directly (data not shown). Thus, PP-1cas likely interacts with p53 indirectly in vivo, a condition necessary for PP-1 to dephosphorylate p53. PP-1 dephosphorylates p53 in vitro Since PP-1cas and p53 form complex in vivo, we next explored whether PP-1 can directly dephosphorylate p53 at Ser-15 and Ser-37. First, we conducted in vitro dephosphorylation assay. To do so, we prepared four

Figure 1 Inhibition of PP-1 (and PP-2A) by okadaic acid and calyculin A leads to induction of apoptosis in HLECs. (A) Activity of PP-1 and PP-2A after treatment with different concentrations of okadaic acid, specific PP-1 or PP-2A inhibitors, or a combination of these agents as shown. (A(a)) Demonstration that 20 nM okadaic acid could inhibit PP-2A; (A(b)) demonstration that 100 nM okadaic acid is able to block PP-1 activity; (A(c)) demonstration that a combination of inhibitors for either PP-1 or PP-2A gives similar inhibition of PP-1 or PP-2A by 20 and 100 nM okadaic acid, respectively. (B) Demonstration that 20 and 100 nM okadaic acid can inhibit PP-2A and PP-1, respectively, in the intact HLECs. HLECs cells were grown to 100% confluence, then treated with 0,20 or 100 nM okadaic acid for 3 h. At the end of each treatment, the cells in each condition were washed with PBS, then immediately lysed in extraction buffer in culture plates and their phosphatase activity under each condition is immediately detected as described previously (Li et al., 1998) and expressed against control level (assuming control sample has an activity of 100%). (C) Demonstration that PP-1 can be inhibited by 100 nM okadaic acid in the intact HLECs through detection of RB phosphorylation status. HLECs were treated with 0.01% DMSO for 24 h (C 24 h), 20 nM okadaic acid for 24 h, or 100 nM okadaic acid for 1–24 h. Then, 30 mg of the total proteins extracted from differentially treated samples was separated by 6% polyacrylamide gel electrophoresis, transferred to nitrocellulose membranes and probed with anti-RB as described before (Li et al., 1998). Represented here in (C) are typical results of three independent experiments. RB-p: hyperphosphorylated form of RB. RB: hypophosphorylated form of RB. (D)–(F) Cell flow cytometry analysis. HLECs were grown to 100% confluence and then treated with 0.01% DMSO (D), 20 nM okadaic acid (E), or 100 nM okadaic acid (F) for 12 h, followed by cell flow cytometry analysis as described (Li et al., 2005). Note that distinct differences exist between 20 and 100 nM okadaic acid-treated samples in early apoptosis (lower right quarter, detected by annexin) and later apoptosis (upper right quarter, detected by annexin and propidium iodide). (G) Concentration-dependent apoptosis induced by calyculin A. HLECs were grown to 100% confluence and then treated with 0 (0.01% DMSO), 1, 3, 6, 12 nM calyculin A for 6 h, followed by cell flow cytometry analysis. (H) Temporal apoptosis pattern induced by calyculin A. HLECs were grown to 100% confluence and then treated with 6 nM calyculin A for 0–24 h, followed by cell flow cytometry analysis. (I) DNA fragmentation assay. HLECs were grown to 100% confluence, then treated with either 0.01% DMSO (lane 2) 24 h, or with 6 nM calyculin A for 2 (lane 3), 6 (lane 4), 12 (lane 5) and 24 h (lane 6). Then, the cells were harvested for DNA fragmentation assays as described previously (Li et al., 1995). The 123 bp marker from Gibco BRL was shown in lane 1. All the experiments have been repeated three times and shown here are typical results. Oncogene


3010

Figure 2 Inhibition of PP-1 activity by okadaic acid or calyculin A dramatically enhances the phosphorylation of p53 at Ser-15 and Ser-37, which causes altered expression of its downstream genes, both bax and bcl-2 at the protein level (a and b) and mRNA levels (c and d). HLECs were treated with 0 (0.01% DMSO) 20, 100 and 200 nM okadaic acid (a and c), or with 0 (0.01% DMSO), 1, 5 and 10 nM calyculin A (b and d) for 3 h and then harvested for extraction of total proteins (a and b) and total RNA (c and d). In all, 100 mg of total proteins was used for Western blot analysis as described before (Li et al., 2001b, 2003). Preparation of total RNA and oligo primers, and analysis of RT–PCR were conducted as described previously (Xiang et al., 2000). The quantitative results of the protein levels for p53, Bax and Bcl-2 in relation to the level of b-actin induced by okadaic acid (e) and calyculin A (f) were determined as previously described (Li et al., 2003). Note that when okadaic acid was increased to a concentration to block PP-1, p53 became hyperphosphorylated at Ser-15 and Ser-37, which causes upregulation of Bax and downregulation of Bcl-2 at the mRNA and protein levels.

different substrate proteins (Figure 5a and b): GST–p53 (containing wild-type p53), GST–p53–S37A (p53 with Ser-37 mutated into Ala), GST–p53–S15A (p53 with Ser-15 mutated into Ala) and GST–p53–S15A/S37A (p53 with both Ser-15 and Ser-37 mutated into Ala). These proteins were first phosphorylated by DNA-PK, a kinase phosphorylating p53 at both Ser-15 and Ser-37 Oncogene

(Lees-Miller et al., 1992). Under these conditions, as shown in Figure 5c, GST–p53 were phosphorylated at both Ser-15 and Ser-37, GST–p53–S37A at Ser-15, GST–p53–S15A at Ser-37 and GST–p53–S15A/S37A none. These substrates were individually mixed with PP-1 and the dephosphorylation buffer, and then incubated for 10 min at 301C. The reaction was terminated


3011

Figure 3 Silencing of p53 expression with RNAi desensitizes the apoptosis induced by okadaic acid and calyculin A. (a) The wild type and mutant siRNA oligos used for RNAi to silence p53 expression in HLECs. (b) Western blot analysis. In all, 100 mg of total proteins extracted from HLECs transfected with wild-type siRNA oligos (left lane) or mutant siRNA oligos (middle lane) for 72 h, or from nontransfected HLECs (right lane) was analysed as described in Figure 2. Shown here are typical results of three independent experiments. (c) Viability assays to show that silence of p53 expression with RNAi substantially blocks apoptosis through inhibition of PP-1 by okadaic acid and calyculin A. HLECs with or without p53 expression silenced were subjected to treatment of different concentrations of okadaic acid (0, 20, 50, 75 nM), calyculin A (0, 1, 3, 6 nM) for 12 h, followed by cell flow cytometry analysis as described in Figure 1. The data shown in (c) and (d) are averaged from three independent experiments.

with 20% TCA. After TCA precipitation, the released free 32P was measured. As shown in Figure 5d, the amount of free 32P released from phosphorylated GST– p53 was doubled than that from GST–p53–S37A or GST–p53–S15A, suggesting that PP-1 dephosphorylates both Ser-15 and Ser-37. However, the amount of free 32P released from GST–p53–S15A/S37A was just background. The dephosphorylation reaction could be inhibited by the specific inhibitor for PP-1, PP1-I2 (Figure 5d). To demonstrate the specificity of the above dephosphorylation reaction, we have prepared the 48 kDa GST–aA-crystallin fusion protein and phosphorylated it by PKA kinase (Chiesa and Spector, 1989). When this substrate was used for the same dephosphorylation assay, no significant amount of free 32 P release was observed (Figure 5). Thus, PP-1 directly dephosphorylates p53 in the in vitro assay. PP-1 dephosphorylates p53 in vivo As PP-1cas and p53 form complex in vivo and PP-1cas can directly dephosphorylate p53 in vitro, we next conducted an in vivo dephosphorylation assay, a procedure developed by Ayllon et al. (2000). HLECs were labeled with [32P]orthophosphate (200 mCi/ml) in phosphatase-free MEM and also subjected to UVA irradiation for 2 h. Under UVA irradiation, both Ser-15

and Ser-37 of p53 became phosphorylated (Bode and Dong, 2004). Total proteins were then extracted for immunoprecipitation with two antibodies against phospho-p53 at Ser-15 and Ser-37. The immunoprecipitated protein complexes were first resolved in the SDS– polyacrylamide gel electrophoresis (SDS–PAGE) and the gel was dried and then exposed to X-ray film to show the presence of PP-1cas and p53 in the immunoprecipitated proteins by anti-p53 antibodies (see the Western blot images within Figure 6a and b). Then, the immunoprecipitated protein complexes were separately resuspended in phosphatase reaction buffer and incubated for 30 min at 301C. After TCA precipitation, the supernatant fraction from each reaction was counted to detect the release of free 32P. As shown in Figure 6, the amount of free 32P released indicated the presence of PP-1 activity, which could be inhibited by okadaic acid and the specific inhibitor for PP-1, but hardly by inhibitors specific for PP-2A and PP-2B. As a negative control, the immunoprecipitated sample with anti-actin antibody was also used for the parallel dephosphorylation assay and the amount of free 32P released was just background (Figure 6). To further confirm that the free 32 P was released from labeled phospho-p53 after dephosphorylation assay, we conducted the same dephosphorylation reaction after the immunoprecipitated protein complex was further precipitated by anti-p53 Oncogene


3012

Biotechnology. When PP-1 activity was mostly inhibited due to loss of PP-1cas through RNAi, hyperphosphorylation of p53 at Ser-15 and Ser-37 was observed even without okadaic acid treatment (Figure 8). These results further confirm that PP-1 is responsible for p53 dephosphorylation at the above sites in the intact HLECs.

Figure 4 Reciprocal immunoprecipitation to demonstrate that PP-1cas and p53 form interacting complex. HLECs were treated with 0 (0.01% DMSO) or 100 nM okadaic acid for 3 h. After treatment, the treated cells were harvested for immunoprecipitation with anti-total p53 antibody (a), or anti-PP-1cas antibody (b), or normal serum (c). The immunoprecipitated samples were then subjected to Western blot analysis as described in Figure 2, using antibodies as indicated. The supernatant represents the sample after immunoprecipitation. Shown here are typical results of three independent experiments.

antibody or anti-PP-1cas antibody. In either case, the amount of free 32P released was just background (data not shown). Thus, PP-1 dephosphorylates p53 at Ser-15 and Ser-37 in the intact HLECs. Overexpression of PP-1cas in HLECs through the tet-on system induced dephosphorylation of p53 at Ser-15 and Ser-37 To further confirm that PP-1 is responsible for dephosphorylating p53 at Ser-15 and Ser-37 in HLECs, we have established a Tet-on overexpression system to overexpress PP-1cas. As shown in Figure 7a, HLECs with the Tet-on cassette of expressing PP-1cas could be induced to overexpress PP-1 activity more than 500-fold by doxycyclin from 0 to 5000 ng/ml. Under maximal induction of PP-1 activity, even in the presence of 100 nM okadaic acid, p53 was completely dephosphorylated at Ser-15 and Ser-37 (Figure 7b). These results confirm that PP-1 dephosphorylates p53 at the above sites in the intact HLECs. Silencing PP-1cas in HLECs through RNAi-induced phosphorylation of p53 at Ser-15 and Ser-37 To further confirm that PP-1 dephosphorylates p53 at Ser-15 and Ser-37, we next conducted RNAi to silence expression of the gene encoding PP-1cas. SiRNA oligo for silencing PP-1cas was obtained from Santa Cruz Oncogene

Dephosphorylation of p53 at Ser-15, Ser-37 or both sites by PP-1 decreases its transcriptional activity From the results described above, we concluded that PP-1 dephosphorylates p53 at Ser-15 and Ser-37 in HLECs. Next, we examined whether dephosphorylation of p53 at Ser-15, Ser-37 or both sites would affect its transcriptional activity. To do so, the reporter gene, pBax-chloramphenicol acetyltransferase (CAT), the transfection efficiency control plasmid, pRSV-b-Gal, and each of the constructs expressing either wild type (pCI-53) or mutant p53 imitating dephosphorylation (pCI-p53-S15A, pCI-p53-S37A, pCI-p53-S15A/S37A) or phosphorylation (pCI-p53-S15D, pCI-p53-S37D, pCI-p53-S15D/S37D) were co-transfected into p53/ mouse lens epithelial cells (MLECs) or p53/ PC-3 cells. A parallel control plasmid, pBax(m)-CAT, in which the p53-binding site was mutated, together with the control plasmid pRSV-b-Gal, and each of the wild-type or mutant p53 expression constructs were also co-transfected into these two types of cells. After 48 h, these transfected cells were harvested for detection of expression of either wild-type or mutant p53 proteins, or used for comparison of the reporter gene (CAT) activity driven by the bax promoter in which an authentic p53-binding site was conserved (Miyashita et al., 1994; Miyashita and Reed, 1995). As shown in Figure 9a and b, in both MLECs and PC-3 cells, wildtype and mutant p53 expression constructs yielded comparable levels of proteins. Transient reporter gene assays revealed that the mutant p53 imitating dephosphorylation all displayed decreased transcriptional activity as compared with the wild-type p53 (Figure 9c and d). In contrast, mutant p53 imitating phosphorylation each showed an enhanced transcriptional activity than the wild-type p53 (Figure 9c and d). In MLECs (Figure 9c), the CAT activity was induced about 12-fold by wild-type p53. Under the same condition, 6.6-, 7.2-, 5.8-, 16.8-, 14.9- and 17.5-fold induction was observed by p53-S15A, p53-S37A, p53-S15A/S37A, p53-S15D, p53-S37D and p53-S15D/S37D, respectively. In PC-3 cells (Figure 9d), the CAT activity was induced about 15-fold by wild-type p53. In contrast, 7.2-, 7.8-, 6.6-, 21.6-, 20.2- and 22.8-fold inductions were obtained by p53-S15A, p53-S37A, p53-S15A/S37A, p53-S15D, p53-S37D and p53-S15D/S37D, respectively. These results clearly showed that dephosphorylation of p53 at Ser-15 and Ser-37 displays strong effects on its transcriptional activity as phosphorylation does. Dephosphorylation of p53 at Ser-15, Ser-37 or both sites by PP-1 attenuates apoptosis To further explore the function of p53 dephosphorylation at Ser-15 and Ser-37, we treated the MLECs and


3013

Figure 5 PP-1 dephosphorylates p53 at Ser-15 and Ser-37 in vitro. The four GST proteins used for substrates were designed as shown in (a). Note that the number of phosphorylation sites by DNA-PK is indicated by ‘p’ above the diagram for each fusion protein. (b) The SDS–PAGE gel separation of the fusion proteins of GST–p53–S15A (lane 2), GST–p53–S37A (lane 3), GST–p53–S15A/S37A (lane 4) and GST–p53 (lane 5) before column purification. The molecular weight marker of 19–118 kDa is shown in lane 1. (c) Columnpurified GST (lane 1), GST–p53 (lane 2), GST–p53–S37A (lane 3), GST–p53–S15A (lane 4) and GST–p53–S15A/S37A (lane 5) without (lower panel) or with (top panel) labeling by DNA-PK were separated in SDS–PAGE gel and then directly stained (lower panel) or exposed to X-ray film (top panel). Shown here are typical results of three independent experiments. (d) In vitro dephosphorylation assays by PP-1. The four different substrates labeled with DNA-PK were precipitated by TCA and then equal amount of each protein (determined by counts) was mixed with 1 PP-1 dephosphorylation buffer, 100 ng PP-1 (obtained from Cell Signaling, Inc.) in the absence or presence of specific PP-1 inhibitor, PP1-I2. The mixture was incubated for 10 min at 301C. After the reaction, the released free 32P was counted. The final results were averaged from three independent experiments in which the free 32P released from each dephosphorylation reaction was recorded.

PC-3 cells expressing p53, p53-S15A, p53-S37A or p53S15A/S37A with 5 nM calyculin A or 75 nM okadaic acid. After treatment for 6 h, the apoptosis was analysed in different samples. As shown in Figure 10, the MLECs (Figure 10a) and PC-3 cells (Figure 10b) expressing wild-type 53 were more susceptible to induced apoptosis than the same cells expressing any of the three types of p53 mutants, p53-S15A, p53-S37A or p53-S15A/S37A. In comparison with the cells transfected with wild-type p53, the MLECs transfected with p53-S15A, p53-S37A or p53-S15A/S37A displayed corresponding 25, 21 and 32% less apoptosis after treatment by okadaic acid and similar decreases in apoptosis after calyculin A treatment (Figure 10a). Similar patterns of attenuated apoptosis in mutant p53-transfected cells were observed in PC3 cells. In all, 27, 22 or 35% less apoptosis was

observed in cells transfected with p53-S15A, p53-S37A or p53-S15A/S37A than those transfected with wild-type p53 after okadaic acid treatment (Figure 10b). A similarly attenuated percentage of apoptosis was detected in PC-3 cells with calyculin A action (Figure 10b). As a comparison, the same two types of cells transfected with p53-S15D, p53-S37D and p53-S15D/S37D were also subjected to treatment by okadaic acid and calyculin A. As shown in Figure 10, the MLECs transfected with p53-S15D, p53-S37D or p53-S15D/S37D showed corresponding 23, 18 and 30% more apoptosis than those transfected with wild-type p53 after treatment by okadaic acid and calyculin A (Figure 10a). Also, 25, 20 and 32% more apoptosis was observed in the PC-3 cells transfected with p53-S15D, p53-S37D or p53-S15A/S37D, respectively, than those transfected Oncogene


3014

Figure 6 In vivo dephosphorylation of p53 at Ser-15 (a) and Ser-37 (b) by PP-1. HLECs were labeled with [32P]orthophosphate (200 mCi/ml) in phosphatase-free DMEM for 2 h. Total proteins were extracted for immunoprecipitation with antibodies against phospho-p53 at Ser-15 (a) or Ser-37 (b). The immunoprecipitated protein complexes were mixed with phosphatase reaction buffer and incubated at 301C for 30 min. After TCA precipitation, the supernatant fraction was recovered for counting the release of free 32P in a scintillation counter. Immunoprecipitated sample with anti-actin antibody was also used for the parallel dephosphorylation assay (mock). The dephosphorylation assays were conducted in the absence (Extract) or presence of PP-1 inhibitors, PP1-I2 (Extract þ PP1I2) or 100 nM okadaic acid (Extract þ OA), PP-2A inhibitor, PP2A-I1 (Extract þ PP2A-I1) and PP-2B inhibitor, 250 nM cyclosporine A (Extract þ CSA). The results shown here are averaged from three independent experiments. Note that inhibitors blocking PP-1, but not PP-2A or PP-2B, were able to inhibit the dephosphorylation.

with wild-type p53 in response to okadaic acid and calyculin A (Figure 10b). Thus, dephosphorylation of p53 by PP-1 at Ser-15 and Ser-37 attenuates p53-dependent apoptosis. Discussion In the present study, we have demonstrated the following: (1) inhibition of PP-1 by okadaic acid and calyculin A induces apoptosis of HLECs; (2) inhibition of PP-1 causes strong phosphorylation of p53 at Ser-15 and Ser-37 and altered expression of the downstream genes bcl-2 and bax; (3) in vitro dephosphorylation Oncogene

assay reveals that PP-1 dephosphorylates p53 at Ser-15 and Ser-37; (4) PP-1cas and p53 form in vivo interaction complex; (5) in vivo dephosphorylation assay shows that PP-1 dephosphorylates p53 at the two sites, which is further confirmed by PP-1cas overexpression through Tet-on system and PP-1cas silence via RNAi; (6) Mutations imitating dephosphorylation of p53 at Ser-15, Ser-37 or both sites by PP-1 attenuate the transcriptional activity of p53 and also p53-dependent apoptosis. Together, our results demonstrate that PP-1 can directly dephosphorylate p53 at Ser-15 and Ser-37, which substantially changes its transcriptional activity and proapoptotic ability. Through modulation of p53 function, PP-1 promotes survival (Figure 11).


3015

Figure 8 Silence of PP-1cas expression through RNAi induces hyperphosphorylation of p53 at Ser-15 and Ser-37. HLECs were transfected with PP-1cassiRNAs or control oligos, and the transfected cells were culture for additional 72 h. Thereafter, the cells were treated with 0.01% DMSO or 100 nM okadaic acid for 3 h, then harvested for extraction of total proteins for Western blot analysis with antibodies against PP-1cas, phospho-p53 at Ser-15 or Ser-37 and b-actin. Note that inhibition of PP-1 activity through RNAi lead to p53 hyperphosphorylation at Ser-15 and Ser-37 in HLECs. Shown here are typical results of three independent experiments.

Figure 7 Overexpression of PP-1cas through Tet-on system induces hypophosphorylation of p53 at Ser-15 and Ser-37. (a) The inducibility of the Tet-on system. The PP-1cas-inducible HLECs grown to 80% confluence were induced by doxycyclin from 0 to 5000 ng/ml for 1 h. At the end of induction, the cells were harvested for extraction of total proteins, which were used for detection of PP-1 activity using a kit from New England Biolab. The results shown here are averaged from three independent experiments. (b) Overexpression of PP-1cas induces hypophosphorylation of p53 at Ser-15 and Ser-37. The stable PP-1casinducible HLECs grown to 80% confluence were pre-induced by 1, 50, 500 and 5000 ng/ml doxycyclin for 1 h followed by induction, and also treatment of 100 nM okadaic acid for additional 3 h. At the end, cells were harvested for analysis of PP-1cas expression and phosphorylation status of p53 at Ser-15 and Ser-37 through Western blot analysis as described in Figure 2. Shown here are typical results of three independent experiments.

Dephosphorylation of p53 displays strong impact on its functions Since its discovery, many laboratories have demonstrated that different signals can activate multiple pathways to phosphorylate p53 at one or more of the identified 17 serine/threonine residues and thus lead to profound changes in the functions of p53 (Giaccia and Kastan, 1998; Meek, 1998; Ashcroft et al., 2000; Bode and Dong, 2004). At the p53 N-terminus, there are 10 serine/threonine residues (Ser-6, -9, -15, -20, -33, -37, -46, and Thr-18, -55 and -81) that have been found phosphorylated by approximately a dozen of different kinases. These include ataxia telangiectasia mutated (ATM; Siliciano et al., 1997; Banin et al., 1998; Khanna et al., 1998; Nakagawa et al., 1999; Chao et al., 2000; Saito et al., 2002), ataxia telangiectasia and Rad 3related protein (ATR; Tibbetts et al., 1999), DNAdependent protein kinases (DNA-PK; Lees-Miller et al., 1992), checkpoint kinase 1 (ChK1; Chehab et al., 1999, 2000; Unger et al., 1999; Shieh et al., 2000) and 2 (ChK2; Chehab et al., 1999; Unger et al., 1999; Hirao et al., 2000), casein kinase 1 (CK1; Knippschild et al., 1997; Dumaz et al., 1999; Higashimoto et al., 2000;

Sakaguchi et al., 2000), protein kinase C (PKC; Chernov et al., 2001), mitogen-activated protein kinases (ERK, p38 kinase and JNKs; Fuchs et al., 1998a, b; Bulavin et al., 1999; Buschmann et al., 2000, 2001; Yeh et al., 2001; She et al., 2000, 2002), MAPK-activated kinase 2 (MAPKAPK2; She et al., 2002), the largest component of the transcription complex TFIID (TAF1; Gatti et al., 2000; Li et al., 2004a), homeodomain- interacting protein kinase-2 (HIPK2) (Hofmann et al., 2002) and the transcriptional complex factor IIH (TFIIH; Lu et al., 1997). Phosphorylation of these sites is associated with changes in stability, transcriptional activity and apoptotic ability of p53. Within the C-terminus, p53 is phosphorylated at Ser-315, -378, -389 and -392 by cyclin-dependent kinase (CDK; Blaydes et al., 2001), aurora kinase A (AURKA; Katayama et al., 2004), glycogen synthase kinase-3b (GSK3b; Qu et al., 2004), protein kinase C (PKC; Takenaka et al., 1995), double-stranded RNA-activated protein kinase (PKR; Cuddihy et al., 1999), casein kinase 2 (CK2; Hupp et al., 1992; Keller et al., 2001) and p38 kinase (p38; Huang et al., 1999). Phosphorylation of these sites enhances the in vitro specific DNA-binding activity of p53. In the DNA-binding domain (central region) of p53, Ser-149, Thr-150 and Thr-155 are phosphorylated by COP9 signalosome-associated kinase complex (CSN-K; Bech-Otschir et al., 2001), which changes the stability of p53. Compared with our extensive knowledge of the protein kinases phosphorylating p53, studies on the phosphatases that dephosphorylate p53 are still in the initiation stage. Although numerous studies have suggested that multiple phosphatases, including PP1 (Takenaka et al., 1995), PP2A (Scheidtmann et al., 1991), PP5 (Zuo et al., 1998), Wip1 and Cdc14 (Fiscella et al., 1997; Li et al., 2000), can dephosphorylate p53, there is no definitive assignment of the concrete phosphatase responsible for dephosphorylating the specific serine/threonine sites until recently. Long et al. Oncogene


3016

Figure 9 Dephosphorylation of p53 at Ser-15 and Ser-37 by PP-1 negatively regulates its transcriptional activity, and phosphorylation of p53 at the same sites yields opposite effects on its transcriptional activity. Expression levels of p53, p53-S15A, p53-S37A, p53-S15A/ S37A, p53-S15D, p53-S37D, p53-S15D/S37D in MLECs (a) or PC-3 cells (b). The p53/ MLECs derived from p53/ mice (a) or p53/ PC-3 cells (b) were transfected with wild-type p53 expression construct (pCI-p53) or each of the mutant p53 expression constructs (pCI-p53-S15A, pCI-p53-S37A, pCI-p53-S15A/S37A, pCI-p53-S15D, pCI-p53-S37D, pCI-p53-S15D/S37D), together with a reporter gene construct, pBax-CAT, and a transfection efficacy plasmid, pRSV-b-Gal. After 48 h, the transfected cells were harvested for extraction of total proteins, which were used for analysis of expression of wild type, and mutant p53 proteins as indicated. Shown here are typical results of three independent experiments. Note that the wild-type p53 level is similar to that of mutant p53 proteins in both types of cells. (c, d) Mutations imitating dephosphorylation or phosphorylation of p53 at Ser-15, Ser-37 attenuate or enhance its transactivity. Both p53/ MLECs (c) and p53/ PC-3 cells (d) were transfected and harvested as described in (a) and (b). The harvested cells were used for extraction of enzymes for the detection of the b-galactosidase and CAT activities as described previously (Xiang et al., 2002). As control, the same reporter gene with the p53-binding site mutated, together with an efficacy plasmid, pRSVb-Gal, and the wild-type p53 expression construct (pCI-p53) or each of the mutant p53 expression constructs (pCI-p53-S15A, pCIp53-S37A, pCI-p53-S15A/S37A, pCI-p53-S15D, pCI-p53-S37D, pCI-p53-S15D/S37D) was also transfected into both p53/ MLECs (c) and p53/ PC-3 cells (d). At 48 h after transfection, these cells were also harvested for analysis of the b-galactosidase and CAT activities as described above. The data presented in (c) and (d) are averaged from three independent experiments.

(2002) have reported that PP-1 may dephosphorylate p53 at Ser-392 in neonatal rat cardiomyocyte. Dohoney et al. (2004) have shown that, in response to DNA damage, protein phosphatase-2A is able to dephosphorylate p53 at Ser-37 in Molt-4 cells. We have briefly reported that PP-1 dephosphorylates p53 at Ser-15 to promote survival in HLECs (Li et al., 2004b). A similar result was recently reported in human Wi38 and HEK393 cells by another group (Haneda et al., 2004). In the present study, we provide strong evidence to show that PP-1 can directly dephosphorylate p53 at Ser-15 and Ser-37. First, PP-1cas and p53 can form a complex. Moreover, in vitro labeled p53 at Ser-15 and Ser-37 by DNA-dependent protein kinases in the presence of g-[32P]ATP can be dephosphorylated by purified PP-1. Furthermore, under UVA irradiation, the pulse-chase labeled total proteins, when immunoprecipitated with antibodies against phosphorylated p53 at either Ser-15 Oncogene

or Ser-37, can be dephosphorylated with co-immunoprecipitated PP-1 catalytic subunit. These data indicate that PP-1 can directly dephosphorylate p53 at Ser-15 and Ser-37. This conclusion is further supported by overexpression of PP-1cas. Overexpression of PP-1cas through the Tet-on system causes dephosphorylation of p53 at both Ser-15 and Ser-37. Furthermore, silence of PP-1cas through RNAi leads to p53 hyperphosphorylation at these sites. As PP-1 directly dephosphorylates these sites, inhibition of PP-1 activity by okadaic acid or calyculin A can cause p53 hyperphosphorylation at Ser-15 and Ser-37, as observed in Figure 2a and b. Together, the results from our studies and others indicate that (1) both PP-1 and PP-2A can directly dephosphorylate p53; and (2) the two phosphatases may dephosphorylate p53 at the same residue in different tissues. In the present studies, our results demonstrate that dephosphorylation at Ser-15 and Ser-37 has a strong


3017

Figure 10 Dephosphorylation of p53 at Ser-15 and Ser-37 by PP-1 attenuates its apoptotic ability. Both p53/ MLECs (a) and PC-3 cells (b) were transfected with pCI-neo vector, wild-type p53 expression construct (pCI-p53), or each of the mutant p53 expression constructs (pCI-p53-S15A, pCI-p53-S37A, pCI-p53S15A/S37A, pCI-p53-S15D, pCI-p53-S37D, pCI-p53-S15D/S37D), as described in Figure 9. Ninety-six hours after transfection, these cells were treated with 75 nM OA or 6 nM calyculin A (CA) for 6 h. After treatment, the different cell samples were harvested for cell viability assays as described previously (Li et al., 1998). The data presented in (c) and (d) are averaged from three independent experiments.

impact on p53 function. First, dephosphorylation is as important as phosphorylation on the transcriptional activity of p53. Mutations imitating dephosphorylation (from serine to alanine) or phosphorylation (from serine to aspartic acid) at Ser-15, Ser-37 or both sites cause significant reduction or enhancement in their transcriptional activity, as tested in p53/ MLECs and PC-3 cells. The three mutants imitating dephosphorylation (p53-S15A, p53-S37A and p53-S15A/S37A) attenuate the transactivity of wild-type p53 by about 40–52% in MLECs (Figure 9c), and 48–56%, respectively, in PC-3 cells (Figure 9d). Under similar test conditions, the three mutants imitating constitutive phosphorylation (p53-S15D, p53-S37D and p53-S15D/S37D) enhance the transcriptional activity of wild-type p53 by about 25–45% in MLECs (Figure 9c), and 35–52% in PC-3 cells (Figure 9d). Our results are consistent with a recent study by Dohoney et al. (2004). These authors have shown that dephosphorylation at Ser-37 by PP-2A downregulates nearly half of the transactivity of the wild-type p53 when assayed with a luciferase reporter

Figure 11 Diagram to show the possible molecular mechanisms by which PP-1 promotes survival through dephosphorylation of p53. PP-1 may normally dephosphorylate p53, and thus negatively regulates the transcriptional and apoptotic abilities of p53. When PP-1 is inhibited by different inhibitors (okadaic acid and calyculin A), p53 in lens epithelial cells becomes hyperphosphorylated at several sites, as shown in circle. Increase in p53 phosphorylation at these sites is known to substantially enhance its transcriptional activity, leading to upregulation of the proapoptotic genes (encoding Bax, etc.) and downregulation of the antiapoptotic genes (encoding Bcl-2, etc.), and activation of the apoptotic program. PP-1 also normally dephosphorylates Rb, favoring the formation of Rb–E2F complex. Inhibition of PP-1 enhances phosphorylation of Rb, leading to release of members of the E2F family, which may activate expression of ARF. ARF then may negatively regulate MDM2 and thus attenuate the MDM2–p53 interaction, inducing p53 upregulation.

gene in p53/ MEFs. In another study by Kaeser et al. (2004), a reduced transactivity of p53-S15A was also observed when tested through induction of p21 and MDM2 proteins in p53/ HCT116 cells. Together, these results indicate that dephosphorylation of p53 by PP-1 or PP-2A has a strong impact on the transcriptional activity of p53. Second, our results also demonstrate that dephosphorylation of p53 attenuates its apoptotic ability, and phosphorylation of p53 enhances its apoptotic ability (Figure 10a and b). The mutant p53 imitating dephosphorylation at Ser-15, Ser-37 or both residues attenuates both okadaic acid- and calyculin A-induced apoptosis by about 21–32% in MLECs, and 22–35% Oncogene


3018

in PC-3 cells (Figure 10). This is not surprising considering that the mutant p53 imitating dephosphorylation at these sites substantially attenuates its transactivity. A decrease in the transactivity of p53 would downregulate expression of the proapoptotic genes and, thus, could attenuate p53-dependent apoptosis. On the other hand, mutant p53 imitating phosphorylation at the same sites enhances apoptosis by about 18–30% in MLECs, and 20–32% in PC-3 cells due to the increased transactivity of p53. Together, these results indicate that dephosphorylation of p53 has as important an impact on its function as its phosphorylation. In our studies, we observed that the double mutant p53-S15A/S37A displays stronger effects on p53 functions than its corresponding single mutants (p15-S15A and p53-S37A) (Figures 9 and 10). This is also true about p53-S15D/S37D and its corresponding single mutants (p53-S15D and p53-37D) (Figures 9 and 10). However, the effect of the double mutant on p53 function is not a simple addition of the effects from the two single mutants. Similar results have been observed by others (Dohoney et al., 2004; Kaeser et al., 2004). This may be derived from the fact that phosphorylation or dephosphorylation of p53 at individual residue changes its configuration and thus affects its function. A double mutant could further enhance (in case of serine/threonine to aspartic acid) or attenuate (in case of serine/threonine to alanine) the optimal configuration of the functional p53, but not necessarily produce an effect of simple addition from two individual mutants. It is well established that PP-1 acquires specificity of function by its association with targeting/regulatory proteins (DePaoli-Roach, 2003). Our results reveal that although PP-1cas and p53 can be co-immunoprecipitated, they do not seem to bind to each other in the in vitro assay. Such results suggest that certain targeting/ regulatory subunits of PP-1 mediate the indirect interaction between PP-1case and p53 in vivo. One of such candidates is p53BP2, which binds both p53 and PP-1cas (DePaoli-Roach, 2003). PP-1 promotes survival through regulation of p53 and other targets It has been well documented that inhibition of PP-1 or/and PP-2A by the marine algae toxins, okadaic acid and calyculin A, induces apoptosis of various types of cells (Boe et al., 1991; Ishida et al., 1992; Mellgren et al., 1993; Kiguchi et al., 1994; Weller et al., 1995; Fernandez-Sanchez et al., 1996; Morana et al., 1996; Sheikh et al., 1996; Yan et al., 1997; Li et al., 1998, 2001a; Garcia et al., 2003). However, why inhibition of PP-1 or/and PP-2A leads to apoptosis remains largely unknown. As we discussed above, inhibition of PP-1 in lens epithelial cells, or inhibition of PP-1 or/and PP-2A in other types of cells, could impair their dephosphorylation of p53, leading to hyperphosphorylation of p53, which enhances the proapoptotic ability of p53. Besides the direct effect on p53 through dephosphorylation, inhibition of PP-1 by these inhibitors may also induce upregulation of p53 expression, as observed in Oncogene

rabbit and rat lens epithelial cells (Li et al., 1998, 2001a). How could inhibition of PP-1 lead to p53 upregulation? It is known that PP-1 can directly dephosphorylate Rb (Ludlow et al., 1993). Inhibition of PP-1 induces hyperphosphorylation of Rb (Figure 1C; Li et al., 1998). It is generally accepted that hyperphosphorylation of Rb allows its release of the bound E2F factors (Nevins, 1992; Weinberg, 1995), leading to activation of E2F-responsive genes (Trimarchi and Lees, 2002). One of the targets is the gene encoding the p19ARF, which is a key component of the p19ARF–MDM2–p53 network (Zhang et al., 1998; Zhu et al., 1999; Trimarchi and Lees, 2002). Activation of p19ARF inhibits MDM2, a p53 ubiquitin ligase, enhancing p53 stabilization (Momand et al., 1992; Kubbutat et al., 1997; Pomerantz et al., 1998; Stott et al., 1998; Zhang et al., 1998; Haupt et al., 1999; Honda and Yasuda, 1999; Weber et al., 1999; Llanos et al., 2001; Zindy et al., 2003). This explains why inhibition of PP-1 activity could lead to observed p53 upregulation in rabbit and rat lens epithelial cells (Li et al., 1998, 2001a). In the HLECs, however, the presence of the SV40 large-T antigen induces extremely high level of background p53 expression and the induction by okadaic acid and calyculin A treatment is almost masked (Figure 2). It is well known that the SV40 large-T antigen can inhibit the DNA-binding activity and transcriptional activity of the tumor suppressor p53 (Bargonetti et al., 1992; Mietz et al., 1992). Our demonstration that inhibition of PP-1 activity by okadaic acid and calyculin A induces strong hyperphosphorylation of p53 and also activates its transcriptional activity in SV40 large-T transformed HLECs suggests that SV 40 large-T suppression of p53 may require dephosphorylation of p53 by PP-1 (and PP-2A). In this regard, it is interesting to note that transformation of human diploid fibroblasts and epithelial cells requires inhibition or alteration of PP-2A by SV 40 small T (Yu et al., 2001; Hahn et al., 2002). In addition, transformation requires inactivation of p53 and Rb by complex formation with SV 40 large-T (Yu et al., 2001; Hahn et al., 2002; Walter, 2003). Together, these results suggest that both PP-1 and PP2A are involved in regulation of cell transformation and thus carcinogenesis. Among the various targets, modulation of p53 and Rb function by these phosphatases may be one of the critical steps. In HLECs, we also observed that knockdown of p53 could not completely inhibit okadaic acid- and calyculin A-induced apoptosis (Figure 3). Such results suggest two possibilities. First, the remaining p53 may function to mediate the residual apoptosis. Alternatively, the remaining apoptosis is derived from the changed phosphorylation status of other targets after inhibition of PP-1. The fact that okadaic acid and calyculin A not only induce apoptosis of the p53/ MLECs and PC-3 cells transfected with various types of p53 but also induce a base level of apoptosis in these p53 minus cells transfected with an empty pCI vector (Figure 10) supports the later possibility. Currently, we are investigating the non-p53 and non-Rb targets in HLECs.


3019

In summary, our results demonstrate that the protein phosphatase-1 can directly dephosphorylate multiple residues of p53, which attenuate the transactivity of p53. As a result, PP-1-mediated dephosphorylation of p53 attenuates p53-dependent apoptosis. Thus, part of the molecular mechanism by which PP-1 promotes survival is to negatively modulate the p53 function through dephosphorylation. Of course, PP-1 modulation of other phosphoproteins through dephosphorylation may also play an important role in promoting survival (Klumpp and Krieglstein, 2002).

Materials and methods Chemicals Various molecular biology reagents were purchased from Invitrogen Life Science Laboratories, Gaithersburg, MD, USA, BD Biosciences Clontech, Palo Alto, CA, USA and Promega Biotech, Madison, WI, USA. Various antibodies were obtained from Cell Signaling Inc., Beverly, MA, USA, Santa Cruz Biotechnology, Santa Cruz, CA, USA and BD Biosciences, Palo Alto, CA, USA. The culture medium, and most other chemicals and antibiotics were purchased from Sigma, St Louis, MO, USA, and Invitrogen Life Science Laboratories, Gaithersburg, MD, USA. Cell culture The HLECs were grown in Dulbecco’s modified Eagle’s minimal essential medium (DMEM) (Invitrogen) containing 10% fetal bovine serum as described previously (Xiang et al., 2002). The medium was prepared in ion-exchanged doubledistilled water to give an osmolarity of 30075 mosmol supplemented with 26 mM NaHCO3 and 50 units/ml penicillin and streptomycin. Media and sera were sterilized by filtration through 0.22-mm filters, with pH adjusted to 7.2. All cells were kept at 371C and 5% CO2 gas phase. Preparation of expression constructs The wild-type p53 expression construct pCMV-p53 was kindly provided by Dr John Reed (Miyashita et al., 1994; Miyashita and Reed, 1995). The p53 coding sequence from pCMV-p53 was exercised and then inserted into pCI-neo to construct pCIp53. The mutant p53 expression constructs were created through a two-step PCR-linked in vitro mutagenesis as described previously (Mao et al., 2004). Four pairs of oligos used for generation of each mutant p53 expression construct are described below: for p53-S15A (pCI-p53-S15A and pGEX4T-p53-S15A), 50 -ATCGAATTCATGGAGGAGCCGCAGT CA-30 and 50 -TGAAAATGTTTCCTGAGCC-AGAGGGGG CTCG-30 , and 50 -GAGCCCCCTCTGGCTCA-GGAAACAT TTTCA-30 and 50 -ACGCTCGAGTCAGTCTGAGTCAGG-30 ; for p53-S37A (pCI-p53-S37A and pGEX-4T-p53-S37A), 50 -AT CGAATTCATGGAGGAGCCGCAGTCA-30 and 50 -ATCC ATTGCTTGGGCCGGCAAGGGGGA-30 , and 50 -TCCCCC TTGCCGGCCC-AAGCAATGGAT-30 and 50 -ACGCTCGA GTCAGTCTGAGTCAGGCCC-30 . For p53-S15A/S37A double mutant (pCI-p53-S15A/S37A and pGEX-4T-p53S15A/S37A), p53-S15A was used as the initial template, and the same oligo pairs used for creation of p53-S37A were utilized for creation of the p53-S15A/S37A. For p53-S15D (pCI-p53-S15D), 50 -ATCGAATTCATGGAGGAGCCGCA GTCA-30 and 50 -AATGTTTCCTGAT-CCAGAGGGGGC TC-30 , and 50 -GAGCCCCCTCTGGATCAGGAAACATT-30 and 50 -ACGCTCGAGTCAGTCTGAGTCAGG-30 ; for p53-

S37D (pCI-p53-S37D), 50 -ATCGAATTCATGGAGGAGCC GCAGTCA-30 and 50 -ATCCATTGCTTGGTC-CGGCAAG GGGG-30 , and 50 -CCCCCTTGCCGGACCAAGCAATGGA T-30 and 50 -ACGCTCGAGTCAGTCTGAGTCAGGCCC-30 . For p53-S15D/S37D double mutant (pCI-p53-S15D/S37D), p53-S15D was used as the initial template, and the same oligo pairs used for creation of p53-S37D were utilized for creation of the p53-S15D/S37D. Transfection of p53/ MLECs and PC-3 cells The p53/ MLECs were derived from p53/ mouse (Donehower et al., 1992; Taconic, Inc.) and p53/ PC-3 cancer cells (Kaighn et al., 1979) were obtained from ATCC. Both types of cells were cultured in DMEM with 10% fetal bovine serum. The mammalian expression vector, pCI-neo, and the wild-type and mutant p53 expression constructs (pCIp53, pCI-p53-S15A, pCI-p53-S37A, pCI-p53-S15A/S37A, pCI-p53-S15D, pCI-p53-S37D and pCI-p53-S15D/S37D) were amplified in DH-5a and purified as described previously (Li et al., 1995). Both wild-type and mutant p53 cDNAs were individually introduced into p53/ MLECs or p53/ PC-3 cells using lipofectomine 2000 agent as described before (Feng et al., 2004; Mao et al., 2004). Treatment by okadaic acid and calyculin A Parental or various transfected HLE cells were grown to 100% confluence. Then, 10 ml of DMEM medium containing 0.01% DMSO (control), 20–200 nM okadaic acid and 1–12 nM calyculin A dissolved in DMSO (the final concentration of DMSO was 0.01%) was used to replace the culture medium and the treatment was continued for various lengths of time as indicated in the figure legends. RNA interference to silence expression of p53 and PP-1cas RNA interference (RNAi) was conducted as described previously (Lassus et al., 2002). The siRNA oligos for 53 was synthesized by the Oligoengine, Inc. or purchased from Santa Cruz Biotechnology. The siRNA oligos for PP-1cas was purchased from Santa Cruz Biotechnology. Transfection was conducted according to the established protocol (Lassus et al., 2002). Briefly, HLECs were seeded in six-well plates and grown to 50% confluence at the time of transfection. At 1 day before transfection, the cells in six-well plates are incubated overnight at 371C and 3% CO2. On the day of transfection, a set of siliconized microfuge tubes were set up for 20 mM siRNA and another set of tubes for lipofectomine 2000 (Invitrogen). For preparation of the transfection cocktail for each culture well, 10 ml of each 20-mM siRNA (either control or experimental siRNAs) was gently mixed with 200 ml of OptiMEM (Invitrogen) in one set of siliconized microfuge tubes. At the same time, 10 ml of lipofectomine 2000 is gently mixed with 50 ml of OptiMEM in another of tubes. After 5 min, the contents in both tubes were combined by gentle pipetting and the transfection mixtures are allowed to sit for 20 min at room temperature. Meanwhile, the cells were rinsed with PBS (Sigma) and 2 ml of growth medium (MEM with 10% fetal bovine serum) was added to each well. The transfection mixtures were then added to each well dropwise, while the plate is gently agitated. The cells were incubated at 371C and 3% CO2 overnight. After 8 h, the cells in each well were replaced with 2 ml of fresh growth medium. At 72 h after transfection, the cells transfected with control and experimental siRNA were harvested separately for extraction of total protein and used for Western blot analysis using an anti-p53 or anti-PP-1cas monoclonal antibodies as described below. Oncogene


3020 For analysis of the response to okadaic acid and calyculin A, the HLE cells with or without p53 silenced were treated with 0.01% DMSO (control), 20–75 nM okadaic acid or 1–6 nM calyculin A for 12 h; the viability of these samples were analysed with cell flow cytometry, Hoechst staining and DNA fragmentation as described previously (Li and Spector, 1996; Li et al., 1998, 2001a, 2005). Protein preparation and Western blot analysis Western blot analysis of total proteins from HLECs, MLECs and PC-3 cells without or with transfection were conducted as described previously (Li et al., 2001b, 2003, 2005; Mao et al., 2001). In vitro binding assay of PP-1cas and p53 In all, 10 ng of PP-1cas (Cell Signaling, Inc.) was immobilized into the nitrocellulose filter and then incubated at 371C with 20 ng of [32P]GST–p53 in the absence or presence of 500 ng cold GST–p53. The filter was then exposed to X-ray film. An interaction assay for PP-1cas and [32P]Bcl-2 was also conducted as a positive control. In vitro phosphorylation of the p53 fusion proteins The fusion proteins of GST–p53, GST–p53–S15A, GST–p53– S37A and GST–p53–S15A/S37A were prepared as described previously (Mao et al., 2004) and phosphorylated by DNAdependent protein kinase (DNA-PK). This kinase was purchased from Promega Inc. and specifically phosphorylates p53 (GST–p53) at Ser-15 and Ser-37 (Lees-Miller et al., 1992). Equal amounts of GST–p53, GST–p53–S15A, GST–p53– S37A and GST–p53–S15A/S37A were each incubated with DNA-PK in the presence of g-[32P]ATP for 30 min at 301C. The reaction was stopped by adding an equal amount of 2 SDS sample buffer if the reaction mixture was used for SDS–PAGE analysis, or by adding 1/5 volume of 100% trichloroacetic acid to precipitate the reaction mixture, which was then dissolved into 40 ml 1 protein phosphatase assay buffer for dephosphorylation assay (Li et al., 1998). aA-crystallin was also labeled with PKA kinase in the presence of g-[32P]ATP for 30 min at 301C and used for control. In vitro dephosphorylation assays The in vitro dephosphorylation assays were conducted in the following manner. To each substrate tube containing GST– p53, GST–p53–S15A, GST–p53–S37A, GST–p53–S15A/S37A or aA-crystallin dissolved in 40 ml 1 PP-1 assay buffer, 100 ng of PP-1 (from Cell Signaling, Inc.) with 5 ml (8 nM) of PP-1 inhibitor, PP1-I2 (specifically inhibiting protein phosphatase 1, IC50 ¼ 2 nM, Park et al., 1994) or with 5 ml 1 PP-1 assay buffer were added. The mixture was incubated for 10 min at 301C and followed by addition of 140 ml 20% TCA and 10 min incubation on ice. The mixture was then centrifuged at 10 000 g at 41C for 15 min, and 150 ml of the supernatant was withdrawn for counting the released free 32P. In vivo dephosphorylation assays The in vivo dephosphorylation assays were conducted as described before (Ayllon et al., 2000). Briefly, HLECs were labeled with [32P]orthophosphate (200 mCi/ml) in phosphatasefree DMEM and subjected to UVA irradiation for 2 h (UVA irradiation induces p53 phosphorylation at both Ser-15 and Ser-37 residues). Then, total proteins were extracted for immunoprecipitation with antibodies against phospho-p53Ser-15 and phospho-p53-Ser-37. The immunoprecipitated protein complex was mixed with phosphatase reaction buffer and incubated at 301C for 30 min. After TCA precipitation, the Oncogene

supernatant fraction was recovered for counting the release of free 32P in a scintillation counter. The immunoprecipitated sample with anti-actin antibody was also used for the parallel dephosphorylation assay (mock). The dephosphorylation assay was conducted in the absence or presence of PP-1 inhibitors, PP1-I2 or 100 nM okadaic acid as described above. Overexpression of PP-1cas through tet-on expression system The Tet-on expression system was established with a kit from BD Biosciences Clontech, Inc. as described previously (Gossen et al., 1995). The pTet-on plasmid was introduced into HLE and stable clones were selected by G418 medium. The PP-1cas cDNA was inserted into the pTRE vector. This construct was verified by sequencing. The pTRE-PP-1cas plasmid and pTKHyg, the plasmid providing hygromycin resistance, were cotransfected into the stable clone of HLE containing the pTeton plasmid. Double selections were conducted with G418 and hygromycin to obtain the inducible clone. Induction of expression was performed using doxycyclin, which is 100 times more potent than tetracycline in the Tet-on induction system according to Gossen et al. (1995). Analysis of transient gene expression For reporter gene activity, the construct of CAT reporter gene driven by the pBax promoter containing a wild-type or mutant p53-binding site (Miyashita et al., 1994; Miyashita and Reed, 1995), together with the control construct expressing b-galactosidase, and wild-type p53 expression construct (pCI-p53) or each of the mutant p53 expression constructs (pCI-p53-S15A, pCI-p53-S37A, and pCI-p53-S15A/S37A, pCI-p53-S15D, pCI-p53-S37D and pCI-p53-S15D/S37D) was introduced into p53 / MLECs and p53/ PC-3 cells. The transfected cells were plated in 100-mm culture dishes for growth and then harvested after 48 h for detection of expression of various forms of p53 and for assays of b-galactosidase and CAT activities as described previously (Xiang et al., 2002). Statistical analysis In the present study, all the data presented are derived from at least three experiments. During data analysis, statistical analysis was conducted for all sets of data when necessary. Both average and standard deviation were calculated and included in the figures.

Abbreviations CAT, chloramphenicol acetyltransferase; DMEM, Dulbecco’s Eagle’s modified minimal essential medium; PMSF, phenylmethylsulfonyl fluoride; HLECs, human lens epithelial cells; MLECs, mouse lens epithelial cells; PP-1, protein serine/ threonine phosphatase-1; PP-1cas, the catalytic subunit of PP1; SDS–PAGE, SDS–polyacrylamide gel electrophoresis; TBS, tris-buffered saline; TBST, tris-buffered saline with tween-20. Acknowledgements We thank Dr John Reed for the pCMV-p53 and pBax-CAT expression constructs; Joram Piatigorsky for pRSV-b-galactosidase expression construct; Dr Venkat Reddy for HLE cell line. This work is supported in part by the NIH/NEI grant EY15765, the Hormel Foundation and the Lotus Scholar Program Funds from Hunan Province Government and Hunan Normal University.


3021 References Ashcroft M, Taya Y, Vousden KH. (2000). Mol Cell Biol 20: 3224–3233. Ayllon V, Martinez AC, Garcia A, Cayla X, Rebollo A. (2000). EMBO J 19: 2237–2246. Bakkenist CJ, Kastan MB. (2003). Nature 421: 499–506. Banin S, Moyal L, Shieh S, Taya Y, Anderson CW, Chessa L et al. (1998). Science 281: 1674–1677. Bargonetti J, Manfredi JJ. (2002). Curr Opin Oncol 14: 86–91. Bargonetti J, Reynisdottir I, Friedman PN, Prives C. (1992). Genes Dev 6: 1886–1898. Bastians H, Ponstingl H. (1994). J Cell Sci 109: 2865–2874. Bech-Otschir D, Kraft R, Huang X, Henklein P, Kapelari B, Pollmann C et al. (2001). EMBO J 20: 1630–1639. Blaydes JP, Luciani MG, Pospisilova S, Ball HM, Vojtesek B, Hupp TR. (2001). J Biol Chem 276: 4699–4708. Bode AM, Dong ZG. (2004). Nat Rev Cancer 4: 793–805. Boe R, Gjertsen BT, Vintermyr OK, Houge G, Lanotte M, Doskeland SO. (1991). Exp Cell Res 195: 237–246. Brewis ND, Street AJ, Prescott AR, Cohen PTW. (1993). EMBO J 12: 987–996. Buckbinder L, Talbott R, Velasco-Miguel S, Takenaka I, Faha B, Seizinger BR et al. (1995). Nature 377: 646–649. Bulavin DV, Saito S, Hollander MC, Sakaguchi K, Anderson CW, Appella E et al. (1999). EMBO J 18: 6845–6854. Buschmann T, Adler V, Matusevich E, Fuchs SY, Ronai Z. (2000). Cancer Res 60: 896–900. Buschmann T, Potapova O, Bar-Shira A, Ivanov VN, Fuchs SY, Henderson S et al. (2001). Mol Cell Biol 21: 2743–2754. Chao C, Saito S, Anderson CW, Appella E, Xu Y. (2000). Proc Natl Acad Sci USA 97: 11936–11941. Chehab NH, Malikzay A, Appel M, Halazonetis TD. (2000). Genes Dev 14: 278–288. Chehab NH, Malikzay A, Stavridi ES, Halazonetis TD. (1999). Proc Natl Acad Sci USA 96: 13777–13782. Chen MX, McPartlin AE, Brown L, Chen YH, Baker HM, Cohen PTW. (1994). EMBO J 12: 4278–4290. Chernov MV, Bean LJ, Lerner N, Stark GR. (2001). J Biol Chem 276: 31819–31824. Chiesa R, Spector A. (1989). Biochem Biophys Res Commun 162: 1494–1501. Chinkers M. (1994). Proc Natl Acad Sci USA 91: 11075–11079. Cohen P. (1989). Annu Rev Biochem 58: 453–508. Cuddihy AR, Wong AH, Tam NW, Li S, Koromilas AE. (1999). Oncogene 18: 2690–2702. DePaoli-Roach AA. (2003) In: Bradshaw RA, Dennis EA (eds). Handbook of Cell Signaling, vol. 1. Elsevier Science Press: San Diego, CA, USA, pp 613–619 (Chapter 103). Dohoney KM, Guillerm C, Whiteford C, Elbi C, Lambert PF, Hager GL et al. (2004). Oncogene 23: 49–57. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery Jr CA, Butel J et al. (1992). Nature (London) 356: 215–221. Dumaz N, Milne DM, Meek DW. (1999). FEBS Lett 463: 312–316. El-Deiry WS. (1998). Semin Cancer Biol 8: 345–357. Feng H, Xiang H, Mao Y-W, Wang J, Liu J-P, Huang X-Q et al. (2004). Oncogene 23: 7310–7321. Fernandez-Sanchez M, Garcia-Rodriguez A, Diaz-Trelles R, Novelli A. (1996). FEBS Lett 398: 106–112. Fiscella M, Zhang H, Fan S, Sakaguchi K, Shen S, Mercer WE et al. (1997). Proc Natl Acad Sci USA 94: 6048–6053. Franklin RA, McCubrey JA. (2000) 14: 2019–2034. Fuchs SY, Adler V, Buschmann T, Yin Z, Wu X, Jones SN et al. (1998a). Genes Dev 12: 2658–2663.

Fuchs SY, Adler V, Pincus MR, Ronai Z. (1998b). Proc Natl Acad Sci USA 95: 10541–10546. Garcia A, Cayla X, Guergnon J, Dessauge F, Hospital V, Rebollo MP et al. (2003). Biochimie 85: 721–726. Gatti A, Li HH, Traugh JA, Liu X. (2000). Biochemistry 39: 9837–9842. Giaccia A, Kastan MB. (1998). Genes Dev 12: 2973–2983. Gjertsen BT, Doskeland SO. (1995). Biochim Biophys Acta 1269: 187–199. Gossen M, Freundlieb S, Bender G, Muller G, Hillen W, Bujard H. (1995). Science 268: 1766–1769. Hahn WC, Dessain SK, Brooks MW, King JE, Elenbaas B, Sabatini DM et al. (2002). Mol Cell Biol 22: 2111–2123. Haneda M, Kojima E, Nishikimi A, Hasegawa T, Nakashima I, Isobe K. (2004). FEBS Lett 567: 171–174. Haupt Y, Maya R, Kazaz A, Oren M. (1999). Nature 387: 296–299. Higashimoto Y, Saito S, Tong XH, Hong A, Sakaguchi K, Appella E et al. (2000). J Biol Chem 275: 23199–23203. Hirao A, Kong YY, Matsuoka S, Wakeham A, Ruland J, Yoshida H et al. (2000). Science 287: 1824–1827. Hofmann TG, Moller A, Sirma H, Zentgraf H, Taya Y, Droge W et al. (2002). Nat Cell Biol 4: 1–10. Honda R, Yasuda H. (1999). EMBO J 18: 22–27. Huang C, Ma WY, Maxiner A, Sun Y, Dong Z. (1999). J Biol Chem 274: 12229–12235. Huang X, Honkananen RE. (1998). J Biol Chem 273: 1462–1468. Hunter T, Karin M. (1992). Cell 70: 375–387. Hupp TR, Meek DW, Midgley CA, Lane DP. (1992). Cell 71: 875–886. Ishida Y, Furukawa Y, De Caprio JA, Saito M, Griffin JD. (1992). J Cell Physiol 150: 484–492. Ishihara H, Martin BL, Brautigan DL, Karaki H, Ozaki H, Kato Y et al. (1989). Biochem Biophys Res Commun 159: 871–877. Israeli D, Tessler E, Haupt Y, Elkeles A, Wilder S, Amson R et al. (1997). EMBO J 16: 4384–4392. Kaeser MD, Pebernard S, Iggo RD. (2004). J Biol Chem 279: 7598–7605. Kaighn ME, Narayan KS, Ohnuki Y, Lechner JF, Jones LW. (1979). Invest Urol 17: 16–23. Katayama H, Sasai K, Kawai H, Yuan ZM, Bondaruk J, Suzuki F et al. (2004). Nat Genet 36: 55–62. Keller DM, Zeng X, Wang Y, Zhang QH, Kapoor M, Shu H et al. (2001). Mol Cell 7: 283–292. Khanna KK, Keating KE, Kozlov S, Scott S, Gatei M, Hobson K et al. (1998). Nat Genet 20: 398–400. Kiguchi K, Glesne D, Chubb CH, Fujiki H, Huberman E. (1994). Cell Growth Differ 5: 995–1004. Klumpp S, Krieglstein J. (2002). Curr Opin Pharmacol 2: 458–462. Knippschild U, Milne DM, Campbell LE, DeMaggio AJ, Christenson E, Hoekstra MF et al. (1997). Oncogene 15: 1727–1736. Kubbutat MH, Jones SN, Vousden KH. (1997). Nature 387: 299–303. Lassus P, Rodriguez J, Lazebnik Y. (2002). Sci STKE 147: PL13. Lees-Miller SP, Sakaguchi K, Ullrich SJ, Appella E, Anderson CW. (1992). Mol Cell Biol 12: 5041–5049. Li HH, Li AG, Sheppard HM, Liu X. (2004a). Mol Cell 13: 867–878. Li L, Ljungman M, Dixon JE. (2000). J Biol Chem 275: 2410–2414. Oncogene


3022 Li DW-C, Fass U, Huizar I, Spector A. (1998). Eur J Biochem 257: 351–361. Li W-C, Kuszak JR, Dunn K, Wang RR, Ma W, Wang GM et al. (1995). J Cell Biol 130: 169–181. Li DW-C, Liu J-P, Wang J, Mao Y-W, Hou L-H. (2003). Invest Ophthalmol Vis Sci 44: 5277–5286. Li DW-C, Liu J-P, Schmid PC, Schlosser R, Feng H. (2004b). Invest Ophthalmol Vis Sci 45(Suppl): E401 (abstract). Li DW-C, Liu J-P, Xiang H, Mao YW, Wang J, Ma WY et al. (2005). Mol Cell Biol 16: 4437–4453. Li W-C, Spector A. (1996). Free Radic Biol Med 20: 301–311. Li DW-C, Xiang H, Fass U, Zhang X-Y. (2001b). Invest Ophthalmol Vis Sci 42: 2603–2609. Li DW-C, Xiang H, Mao YW, Wang J, Fass U, Zhang XY et al. (2001a). Exp Cell Res 266: 279–291. Llanos S, Clark PA, Rowe J, Peters G. (2001). Nat Cell Biol 3: 445–452. Long X, Wu G, Gaa ST, Rogers TB. (2002). Apoptosis 7: 31–39. Lu H, Fisher RP, Bailey P, Levine AJ. (1997). Mol Cell Biol 17: 5923–5934. Ludlow JW, Glendening CL, Livingston DM, De Caprio JA. (1993). Mol Cell Biol 13: 367–372. Manfredi JJ. (2003). Mol Cell 11: 552–554. Mao YW, Liu J-P, Xiang H, Li DW. (2004). Cell Death Differ 11: 512–526. Mao YW, Xiang H, Wang J, Korsmeyer S, Reddan J, Li DW. (2001). J Biol Chem 276: 43435–43445. Meek DW. (1998). Cell Signal 10: 159–166. Mellgren G, Vintermyr OK, Boe R, Doskeland SO. (1993). Exp Cell Res 205: 293–301. Mietz JA, Unger T, Huibregtse JM, Howley PM. (1992). EMBO J 11: 5013–5020. Mihara M, Erster S, Zaika A, Petrenko O, Chittenden T, Pancoska P et al. (2003). Mol Cell 11: 577–590. Miyashita T, Krajewski S, Krajewska M, Wang HG, Lin HK, Liebermann DA et al. (1994). Oncogene 9: 1799–1805. Miyashita T, Reed JC. (1995). Cell 80: 293–299. Momand J, Zambetti GP, Olson DC, George D, Levine AJ. (1992). Cell 69: 1237–1245. Morana SJ, Wolf CM, Li J, Reynolds JE, Brown MK, Eastman A. (1996). J Biol Chem 271: 18263–18271. Mumby MC, Walter G. (1993). Physiol Rev 73: 673–699. Nakagawa K, Taya Y, Tamai K, Yamaizumi M. (1999). Mol Cell Biol 19: 2828–2834. Nevins JR. (1992). Science 258: 424–429. Oren M. (2003). Cell Death Differ 10: 431–442. Park I, Roach P, Bondor J, Fox SP, DePaoli-Roach AA. (1994). J Biol Chem 269: 944–954. Pawson T. (1995). Nature 373: 573–579. Polyak K, Xia Y, Zweier JL, Kinzler KW, Vogelstein B. (1997). Nature 389: 300–305. Pomerantz J, Schreiber-Agus N, Liegeois NJ, Silverman A, Alland L, Chin L et al. (1998). Cell 92: 713–723. Qu L, Huang S, Baltzis D, Rivas-Estilla AM, Pluquet O, Hatzoglou M et al. (2004). Genes Dev 18: 261–277. Sakaguchi K, Saito S, Higashimoto Y, Roy S, Anderson CW, Appella E. (2000). J Biol Chem 275: 9278–9283.

Oncogene

Saito S, Goodarzi AA, Higashimoto Y, Noda Y, Lees-Miller SP, Appella E et al. (2002). J Biol Chem 277: 12491–12494. Scacheri PC, Rozenblatt-Rosen O, Caplen NJ, Wolfsberg TG, Umayam L, Lee JC et al. (2004). Proc Natl Acad Sci USA 101: 1892–1897. Scheidtmann KH, Mumby MC, Rundell K, Walter G. (1991). Mol Cell Biol 11: 1996–2003. Schular M, Green DR. (2005). Trends Genet 21: 182–187. She QB, Chen N, Dong Z. (2000). J Biol Chem 275: 20444–20449. She QB, Ma WY, Dong Z. (2002). Oncogene 21: 1580–1589. Sheikh MS, Garcia M, Zhan Q, Liu Y, Fornace Jr AJ. (1996). Cell Growth Differ 7: 1599–1607. Shen Y, White E. (2001). Adv Cancer Res 82: 55–84. Shieh SY, Ahn J, Tamai K, Taya Y, Prives C. (2000). Genes Dev 14: 289–300. Shieh SY, Ikeda M, Taya Y, Prives C. (1997). Cell 91: 325–334. Siliciano JD, Canman CE, Taya Y, Sakaguchi K, Appella E, Kastan MB. (1997). Genes Dev 11: 3471–3481. Stott FJ, Bates S, James MC, McConnell BB, StaRborg M, Brookes S et al. (1998). EMBO J 17: 5001–5014. Takenaka I, Morin F, Seizinger BR, Kley N. (1995). J Biol Chem 270: 5405–5411. Tibbetts RS, Brumbaugh KM, Williams JM, Sarkaria JN, Cliby WA, Shieh SY et al. (1999). Genes Dev 13: 152–157. Trimarchi JM, Lees JA. (2002). Nat Rev Mol Cell Biol 3: 11–20. Unger T, Juven-Gershon T, Moallem E, Berger M, Vogt Sionov R, Lozano G et al. (1999). EMBO J 18: 1805–1814. Vousden KH, Lu X. (2002). Nat Rev Cancer 2: 594–604. Walter G. (2003) In: Bradshaw RA, Dennis EA (eds). Handbook of Cell Signaling, vol. 1. Elsevier Science Press: San Diego, CA, USA, pp 621–626 (Chapter 104). Weber JD, Taylor LJ, Roussel MF, Sherr CJ, Bar-Sagi D. (1999). Nat Cell Biol 1: 20–26. Weller M, Malipiero U, Groscurth P, Fontana A. (1995). Exp Cell Res 221: 395–403. Wen-Schaub LB, Zhang W, Cusack JC, Angelo LS, Santee SM, Fujiwara T et al. (1995). Mol Cell Biol 15: 3032–3040. Weinberg RA. (1995). Cell 81: 323–330. Wu GS, Burns TF, McDonald III ER, Jiang W, Meng R, Krantz ID et al. (1997). Nat Genet 17: 141–143. Xiang H, Wang J, Mao Y, Li DW. (2000). Biochem Biophys Res Commun 278: 503–510. Xiang H, Wang J, Mao Y, Liu M, Reddy VN, Li DW. (2002). Oncogene 21: 3784–3791. Yan Y, Shay JW, Wright WE, Mumby MC. (1997). J Biol Chem 272: 15220–15226. Yeh PY, Chuang SE, Yeh KH, Song YC, Cheng AL. (2001). Biochem Biophys Res Commun 284: 880–886. Yu J, Boyapati A, Rundell K. (2001). Virology 290: 192–198. Zhang Y, Xiong Y, YaRbrough WG. (1998). Cell 92: 725–734. Zhu JW, DeRyckere D, Li FX, Wan YY, DeGregori J. (1999). Cell Growth Differ 10: 829–838. Zindy F, Williams RT, Baudino TA, Rehg JE, Skapek SX, Cleveland JL et al. (2003). Proc Natl Acad Sci 100: 15930–15935. Zuo Z, Dean NM, Honkanen RE. (1998). J Biol Chem 273: 12250–12258.